Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18;
Date:   Fri, 9 Oct 2020 15:01:45 -0300
From:   Jason Gunthorpe <jgg@ziepe.ca>
To:     Daniel Vetter <daniel.vetter@ffwll.ch>
Cc:     Mauro Carvalho Chehab <mchehab+huawei@kernel.org>,
        DRI Development <dri-devel@lists.freedesktop.org>,
        LKML <linux-kernel@vger.kernel.org>,
        KVM list <kvm@vger.kernel.org>, Linux MM <linux-mm@kvack.org>,
        Linux ARM <linux-arm-kernel@lists.infradead.org>,
        linux-samsung-soc <linux-samsung-soc@vger.kernel.org>,
        "open list:DMA BUFFER SHARING FRAMEWORK" 
        <linux-media@vger.kernel.org>,
        linux-s390 <linux-s390@vger.kernel.org>,
        Daniel Vetter <daniel.vetter@intel.com>,
        Kees Cook <keescook@chromium.org>,
        Dan Williams <dan.j.williams@intel.com>,
        Andrew Morton <akpm@linux-foundation.org>,
        John Hubbard <jhubbard@nvidia.com>,
        =?utf-8?B?SsOpcsO0bWU=?= Glisse <jglisse@redhat.com>,
        Jan Kara <jack@suse.cz>,
        Linus Torvalds <torvalds@linux-foundation.org>
Subject: Re: [PATCH v2 09/17] mm: Add unsafe_follow_pfn
Message-ID: <20201009180145.GB5177@ziepe.ca>
References: <20201009075934.3509076-1-daniel.vetter@ffwll.ch>
 <20201009075934.3509076-10-daniel.vetter@ffwll.ch>
 <20201009123421.67a80d72@coco.lan>
 <20201009122111.GN5177@ziepe.ca>
 <20201009143723.45609bfb@coco.lan>
 <20201009124850.GP5177@ziepe.ca>
 <CAKMK7uF-hrSwzFQkp6qEP88hM1Qg8TMQOunuRHh=f2+D8MaMRg@mail.gmail.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <CAKMK7uF-hrSwzFQkp6qEP88hM1Qg8TMQOunuRHh=f2+D8MaMRg@mail.gmail.com>
Precedence: bulk

On Fri, Oct 09, 2020 at 07:52:05PM +0200, Daniel Vetter wrote:

> > > If this is the case, the proper fix seems to have a GFP_NOT_MOVABLE
> > > flag that it would be denying the core mm code to set __GFP_MOVABLE.
> >
> > We can't tell from the VMA these kinds of details..
> >
> > It has to go the other direction, evey mmap that might be used as a
> > userptr here has to be found and the VMA specially created to allow
> > its use. At least that is a kernel only change, but will need people
> > with the HW to do this work.
> 
> I think the only reasonable way to keep this working is:
> - add a struct dma_buf *vma_tryget_dma_buf(struct vm_area_struct *vma);
> - add dma-buf export support to fbdev and v4l
> - roll this out everywhere we still need it.

It seems to me there is a technical way forward to restore user
compat, so it is really no different than RDMA/DRM pain we both
suffered before.

Thus no justification to NAK it. If media wants things to keep working
they have to do the technical path like you outline above.

> Realistically this just isn't going to happen. 

If your series goes ahead it will get solved. Someone will take on the
huge project to either add DMA buf to the drivers people still care
about, or do the work above to transparently handle in kernel.

If we allow things to keep working without consequence then nobody
will do it.

The only reason we did the 4 years of work in RDMA was because Linus
went in and broke the uABI for a security fix. It was hundreds of
patches to fix it, so I don't have much sympathy for "it is too hard"
here.

Jason